115
Chapter 7
Novel
Communications-Based
Train Control System
with Coordinated
Multipoint Transmission
and Reception
Li Zhu, F. Richard Yu, and Tao Tang
Contents
7.1 Introduction .............................................................................................116
7.2 CBTC Systems .........................................................................................119
7.2.1 Impacts of Communication Latency on CBTC Systems ............... 119
7.2.2 Proposed CBTC System with CoMP ............................................121
7.3 System Models .........................................................................................122
7.3.1 Train Control Model ....................................................................123
7.3.2 Communication Channel Model .................................................. 125
7.4 Communication LatencyinCBTCSystems with CoMP .........................126
7.4.1 Coordinated Multipoint Transmission and Reception ..................126
7.4.2 Data Transmission Rate and BER .................................................126
7.4.3 Communication Latency ..............................................................128
116 ◾ Advances in Communications-Based Train Control Systems
7.1 Introduction
In existing train–ground communication systems, the mobile terminal (MT) on a train
communicates with an independent wayside base station to realize train–ground com-
munications. Trains travel fast on the railways, and the received signal-to-noise ratio
(SNR) changes rapidly. e communication latency will be a serious problem when
the MT on a train is in deep fading. More importantly, when a train moves away from
the coverage of a base station and enters the coverage of another base station along the
railway, a hando procedure occurs. In the current communications-based train control
(CBTC) systems, only the traditional hard hando scheme is supported, where a train
can only communicate with a single base station at any given time. It may result in com-
munication interruption and long latency due to the weak wireless signals in the hando
zone and high moving speed of trains. Both of these two challenges can severely aect
train control performance, train operation eciency, and the utilization of railway.
Recently, some research has been done on the train–ground communication
issues in the railway environment. In [1], a fast handover algorithm suitable for
dedicated passenger line is proposed by setting a new neighboring list. A novel
handover scheme based on on-vehicle antennas is introduced in [2]. A cross-layer
hando design is studied in [3] for multiple-input and multiple-output (MIMO)-
enabled wireless local area networks (WLANs).
Although these above works consider the impacts of railway environment on
the communication performance, the hando schemes in most current research
7.5 Control Performance OptimizationinCBTC Systems with CoMP ......... 129
7.5.1 SMDP-Based CoMP Cluster Selection and Hando
DecisionModel ............................................................................129
7.5.1.1 Decision Epochs .............................................................130
7.5.1.2 Actions ...........................................................................130
7.5.1.3 States ..............................................................................131
7.5.1.4 Reward Function ............................................................131
7.5.1.5 State Transition Probability ............................................132
7.5.1.6 Constraints .....................................................................134
7.5.2 Solutions to SMDP-Based CoMP Cluster Selection and
Hando Decision Scheme ............................................................134
7.5.2.1 Reduced-State Bellman Equation ...................................135
7.5.2.2 Online Value Iteration AlgorithmviaStochastic
Approximation................................................................135
7.5.3 Optimal Guidance Trajectory Calculation ...................................137
7.6 Simulation Results and Discussions .........................................................139
7.6.1 Train Control Performance Improvement ..................................... 140
7.6.2 Hando Performance Improvement .............................................143
7.7 Conclusion ............................................................................................... 145
References .........................................................................................................146
Novel Communications-Based Train Control System ◾ 117
are still hard hando based, where the “break-before-make” principle limits the
hando performance improvement. In addition, the impacts of hando latency
on the control performance of CBTC systems is largely ignored in the existing
works.
In this chapter, we use recent advances in coordinated multipoint (CoMP)
transmission/reception to enable soft hando, and consequently enhance the perfor-
mance of CBTC systems. CoMP is a new method that helps with the implementation
of dynamic base station coordination in practice. It is considered as a key technology
for future mobile networks and is expected to be deployed in the future long-term
evolution-advanced (LTE-A) systems to improve the cellular network performance
[4]. To the best of our knowledge, using CoMP in CBTC systems has not been stud-
ied in previous works.
Intuitively, CoMP can improve not only the performance of commercial net-
works (e.g., cellular networks) also the performance of CBTC systems. Nevertheless,
the adoption of CoMP in CBTC is not trivial due to the following reason: Traditional
design criteria, such as network capacity, are used in existing works in CoMP-based
networks. However, recent studies in cross-layer design show that maximizing capac-
ity does not necessarily benet the application layer [5–8], which is train control in
CBTC systems. From a CBTC perspective, the performance of train control is more
important than that at other layers. A commonly used control performance measure
is the linear quadratic cost [9], which is directly related to train control accuracy,
train safety, and passengers ride quality [10].
In this chapter, we propose a CBTC system with CoMP to enhance the train
control performance of this system. e distinct features of this chapter are as follows:
1. We propose a CoMP-enabled CBTC train–ground communication system.
With CoMP, a train can communicate with a cluster of base stations simul-
taneously, which is dierent from the current CBTC systems, where a train
can only communicate with a single base station at any given time.
2. Unlike the existing works on communication systems that use capacity as the
performance measure, in this chapter, linear quadratic cost for the train con-
trol performance in CBTC systems is considered as the performance measure.
3. We jointly consider CoMP cluster selection and hando decision issues in CBTC
systems. Although some works have been done to address the hando prob-
lem, most of them focus on hando protocols, and consequently hando deci-
sion policy issues (i.e., when to perform hando) are largely ignored in CBTC
systems, which should be carefully considered. e hando decision problem
becomes more complicated when CoMP is used in CBTC, because the system
needs to decide not only when to perform hando but also which cluster to use.
4. In order to mitigate the impacts of communication latency on the train con-
trol performance, we propose an optimal guidance trajectory calculation
scheme in the train control procedure that takes full consideration of the
tracking error caused by communication latency.
118 ◾ Advances in Communications-Based Train Control Systems
5. e system optimization of CBTC system with CoMP is formulated as a
semi-Markov decision process (SMDP) [11], which has been successfully used
to solve opportunistic spectrum access in cognitive networks [12], among
others. is chapter focuses on the application of SMDP to the control per-
formance improvement in CBTC systems with CoMP.
6. Extensive simulation results are presented. It is shown that train control perfor-
mance can be improved substantially in our proposed CBTC system with CoMP.
e main notations used in this chapter are summarized in Table 7.1.
Table 7.1 Main Notations
Notation Denition Unit
T Communication period s
q(k) Train position m
v(k) Train velocity m/s
ε(k) Velocity tracking error m/s
M Train Mass kg
w
i
(k) Slope resistance N
w
r
(k) Curve resistance N
w
w
(k) Wind resistance N
u(k) Train control command N
x(k) Train states m, m/s
y(k) Controller input m, m/s
x
c
(k) Controller state m/s
z(k) Observed train states m/s
α
Deceleration of the ATP braking curve m/s
2
C(θ) Channel capacity under cluster θ bits/s
P
o
Transmitted power of each user terminal mW
PR
Td
Channel transition probability after latency T
d
h
B*
(k) Channel gain from a base station to the MT dB
T
l
Current communication latency ms
T
a
Communication latency from the front train to the ground ms
Novel Communications-Based Train Control System ◾ 119
e rest of this chapter is organized as follows: Section 7.2 introduces CBTC
systems. Section 7.3 describes the system models. Section 7.4 discusses the deriva-
tion of communication latency. Section 7.5 describes control performance optimi-
zation. Section 7.6 presents simulation results and discussions. Finally, Section 7.7
concludes this chapter.
7.2 CBTC Systems
In this section, we rst present the impacts of wireless communications on CBTC
systems that will be introduced next. en, we describe the proposed CBTC system
with CoMP.
7.2.1 Impacts of Communication Latency on CBTC Systems
e impacts of wireless communications on CBTC can be categorized into two
aspects: safety and eciency. We rst give a simple example about the safety. Based on
the locations of all the trains and other obstacles along the railway, the zone controller
(ZC) transmits a movement authority (MA) to each train. One kind of obstacles is
the switch. When a train approaches to a switch area, the ZC that controls this area
will send the switch state to the train. When the switch state is in a normal state, the
train will go through the switch to the normal direction without decreasing speed.
However, the switch state may change to reverse or unknown due to equipment fail-
ure or human error. If the wireless link is not available due to communication latency
at this moment, the ZC will not be able to report the updated switch state to the train.
With long communication latency, the train will then go through the switch to the
reversed direction with full speed, and an accident may happen.
Next we present the impacts on CBTC eciency. In every communication period,
the train and ground equipment sends the corresponding information to each other
after processing the received information. As shown in Figure 7.1a, when there is
no communication latency, Train 2 gets the updated MA in every communication
period. e MA moves forward with Train 1. e velocity of Train 2 will not reach
the target velocity on the braking curve when the two trains travel along the same
guidance trajectory. Due to communication latency in train–ground communica-
tion systems, Train 2 may not be able to get the updated MA from ZC for a certain
period of time. As shown in Figure 7.1b, Train 2 will still use the old MA before
the updated MA arrives, and its automatic train protection (ATP) subsystem will
start service brake to protect the train from traveling out of its MA. When the ATP
subsystem receives the updated MA, it stops the service brake. e automatic train
protection (ATO) system will then take control of the train and bring the train
back to the original optimized guidance trajectory.
To make it easier to understand the impact of communication latency on CBTC
eciency, we have added a simulation result. We only simulate the communication
..................Content has been hidden....................
You can't read the all page of ebook, please click here login for view all page.